If you own an Amazon Echo or use Google Maps, you’re already familiar with the benefits of artificial intelligence (AI).
Whether understanding garbled verbal instructions or finding the quickest route downtown, AI is proving incredibly useful in a wide variety of applications. Hiring and recruiting are no exception.
A few weeks ago, HireVue’s Chief I-O Psychologist, Dr. Nathan Mondragon, sat down to explain the implications of artificial intelligence on sourcing and recruiting.
He started with a brief overview of the three levels of AI:
The weakest level of artificial intelligence was developed in the 1990s, like IBM’s Deep Blue Master Chess Player. While there are still a few technologies making use of this variety of AI, most fall solidly into the middle tier.
The “middle” level of AI is where the technology currently stands. Machine learning uses the human reasoning process as a guide, but does not fundamentally replicate it. According to Nathan Mondragon, this is probably a good thing:
“In many cases we don’t want to exactly model the human thinking process, because sometimes there are a lot of bias and errors there.” - Nathan Mondragon
While “AI” has become something of a buzzword of late, artificial intelligence software in its current state has found widespread application in the general populace:
- Amazon’s Alexa: uses AI to understand and learn (often garbled) voice commands, getting better over time.
- IBM’s Watson Health: assists radiologists in distinguishing between malignant and benign tumors - often with higher accuracy than the human radiologists.
- Otto, the self-driving truck: drove 120 miles on a beer delivery through the heart of Denver.
With AI proving useful from the kitchen to the doctor’s office, it’s only natural that it finds application for recruiters.
Applications to Hiring & Recruiting
Mondragon identifies three tiers of AI application emerging in the talent acquisition space. While some do a better job of predicting job performance, each brings its own unique benefits to the table:
That something is a "weak" or "average" predictor isn't necessarily a bad thing. Many of the software tools listed above have uses outside of specifically predicting job performance (like finding passive candidates).
According to Mondragon, most of the activity in the “AI in HR” space falls under the category of “weak predictors.” By scraping data from candidate’s social media, machine learning algorithms can assist recruiters with sourcing and screening tasls.
For example: if a candidate’s LinkedIn profile shows a high level of engagement with their current or previous roles, they are likely to be engaged in their future positions as well. Precisely measuring, quantifying, and predicting that engagement is where the machine learning and AI technology come into play.
Another weak predictor of job performance is the AI-driven chatbot. These engage with candidates (often on social media) and ask them job-related questions, often about their past experience. After gathering this sort of basic information, the AI-driven chatbot is able to make first-pass screening decisions on its own and inform the candidate accordingly.
In the middle are things like simulations, testing, and gamification. Unlike most weak predictors, these collect data directly from the candidate with more structured approach. Tests of this variety are “one-size-fits-all” - they can tell you a person is probably a good hire based on generic traits that are predictive of success (think grit, determination, and tenacity).
The custom AI-driven assessment is, as Mondragon explains, “where the real quality of hire comes in.”
The “strong predictors” are custom assessments built for specific jobs, and unlike the average predictors, are created with specific job metrics in mind. By linking to each job’s unique performance measures, these AI-driven assessments use custom algorithms to connect the dots between assessment data and job success.
Since they are built for specific roles, they can also be vetted for adverse impact, ensuring that the assessment is treating all groups fairly.
“Really there’s only HireVue in this column today - this isn’t really a self-promoting kind of thing for the company. There are others coming and looking at it, but we are really the only company doing this type of work today.” - Nathan Mondragon
Does Hiring Really Need AI?
Is hiring really broken and in need of fixing? According to several CHRO roundtables Mondragon participated in, the answer is a resounding “yes.”
- Communications are poor and candidates are lost in the shuffle. 50-60% of candidates are applying and not hearing back from the hiring companies.
- The effect of the hiring process on job performance is largely unknown. HR doesn’t know what steps in the hiring process are identifying top performers and what steps are screening them out.
- Despite these problems, there isn’t really time to rethink the hiring process - HR is just too busy.
AI has the potential to fix these issues and bring in what Mondragon terms “Hiring Nirvana.”
Entering “Hiring Nirvana”
So what is “Hiring Nirvana”? According to Mondragon, “Hiring Nirvana” upgrades the following three metrics:
- Time to Hire. Efficient, cost-effective hiring practices: getting to talent faster.
- Quality of Hire. Measurable indicators of candidate success: identifying the best talent.
- Quality of Experience. A modern, secure, and user-friendly system for recruiters and candidates: providing a great candidate experience.
Traditionally, talent acquisition needed to sacrifice speed of hire for quality of hire (or vice versa). Each additional round of interviews increases the likelihood of hiring the best fit candidate, but might add weeks to the hiring process.
With AI, speedy hires can be achieved - not at the expense of quality, but with its addition.
How HireVue Achieves "Hiring Nirvana"
At HireVue, our team of Industrial-Organizational (I-O) Psychologists (led by Dr. Nathan Mondragon) leverage video interviewing to deliver custom, “strong predictor” assessments. They create questions specifically designed to elicit responses predictive of job success and find the right behaviors.
With artificial intelligence, candidate’s video responses are scored against a sophisticated model of the position's most desirable competencies and attributes - providing an I-O validated prediction of job success from a candidate friendly talent assessment:
“What we’re doing with video is truly groundbreaking - no one else is doing it. We’re getting these 5-6 day hiring cycles, but we’re also documenting and justifying the ROIs with customers where they have, for example, a reduction in safety violations by 50% or a 31% increase in tenure for call center employees.” - Nathan Mondragon
So with a HireVue Assessment, you’re really combining these hiring process steps into one:
- Resume Review.
- Phone Screen.
- Pre-Employment Assessment Test.
- First Round Interview.
What ends up happening is hiring managers get sent a small number of candidates, all of them are quality - they just choose the one that fits their team the best.
Different AI For Different Hiring Approaches
While this piece focused on HireVue pretty heavily (this is our blog, after all), at the end of the day there are a variety of different solutions for different problems.
If you want to tackle recruiting from the sourcing and job-matching side, there are AI-driven solutions there.
If you want a one-size-fits-all approach that you can use to test for any job, there are solutions there as well.
Of course, if you want a custom approach for your company, jobs, and success metrics that maximizes legal defensibility - that’s possible too.